跳到正文
arXiv cs.AI· Sayumi Muthukumarana, Roshan Godaliyadda, Parakrama Ekanayake, Roshan Ragel·· 10 小时前AI 评分42

LLM 生成的儿童睡前故事存在性别刻板印象:一项覆盖 16 个模型的大规模研究

Once Upon a Bias: Gender Stereotyping in LLM-Generated Children's Stories

AI 导读

一项大规模研究分析了 16 个 LLM 在 35 个国家设定、12 种叙述者角色下生成的 62,873 篇睡前故事,发现面向儿子的故事比面向女儿的故事更强调能动性、更少强调共情。研究采用匹配的女儿/儿子提示词与五次生成设计,结合词汇率对比、相对刻板印象对齐、叙述者-接收者交互分析、人工评估与表征层干预。隐藏状态分析显示接收者性别信息可被解码,作者据此呼吁超越短文本基准、开展叙事层面的偏见评估。

正文 · 原文

View PDF HTML (experimental)

Abstract:Large language models are increasingly used to create stories for children. Since bedtime stories can shape children's perceptions of social roles, this raises concerns about whether LLMs adapt narrative portrayals according to the intended recipient's gender. We conducted a large-scale study to examine gender stereotyping in LLM-generated bedtime stories. Using matched daughter/son prompts and a five-generation design, we analyze 62,873 retained stories from 16 LLMs across 35 country settings and 12 narrator roles. We introduce a multi-level evaluation framework combining context-adjusted lexical rate contrasts, relative stereotype alignment, narrator-recipient interaction analysis, human evaluation, and representation-level interventions to examine both the consistency and interpretability of these differences. Our results reveal systematic differences in narrative framing: son-directed stories emphasize higher agency and lower communion compared to daughter-directed stories. While narrator gender expresses limited moderation on these patterns, high sociability exhibits a more pronounced interaction. Human evaluations demonstrate consistent agreement, although correspondence with lexical measures varies across narrative dimensions. Additionally, hidden-state analyses reveal decodable recipient-gender information. These findings demonstrate persistent yet context-dependent representational disparities in LLM-generated children's stories and motivate narrative-specific bias evaluation beyond short-form benchmarks.
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Cite as: arXiv:2611.24156 [cs.CL]
  (or arXiv:2611.24156v1 [cs.CL] for this version)
  https://doi.org/10.48550/arXiv.2611.24156

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Sayumi Muthukumarana [view email]
[v1] Wed, 18 Nov 2026 10:01:35 UTC (16,087 KB)

来源:arXiv cs.AI · arxiv.org